andrew.sayer(1992)method.in.social.science a.realist.approach.2ed.ed

Method in Social Science
A realist approach

Second Edition

Andrew Sayer

London and New York

First published in 1984 by Hutchinson

Second edition published in 1992
by Routledge
11 New Fetter Lane, London EC4P 4EE

Simultaneously published in the USA and Canada
by Routledge
29 West 35th Street, New York, NY 10001

Routledge is an imprint of the Taylor & Francis Group

This edition published in the Taylor & Francis e-Library, 2003.

© 1992 Andrew Sayer

The author asserts the moral right to be
identified as the author of this work.

All rights reserved. No part of this book may be reprinted or reproduced
or utilized in any form or by any electronic, mechanical, or other means,
now known or hereafter invented, including photocopying and recording,
or in any information storage or retrieval system, without permission in
writing from the publishers.

British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library

Library of Congress Cataloguing in Publication Data
A catalogue record for this book is available from the Library of Congress

ISBN 0-203-16360-5 Master e-book ISBN

ISBN 0-203-16372-9 (Adobe eReader Format)
ISBN 0-415-07607-2 (Print Edition)

Contents

Preface to the second edition page ix

Introduction 1

1 Knowledge in context 12
Some misconceptions about knowledge—Knowledge,
work and communicative interaction—The relationship
between subject and object—Some implications of sub-
ject-object relations—Verstehen—Critical theory and the
relationship between subject and object—Conclusions

2 Theory, observation and practical adequacy 45
Knowledge and object—‘Theory’—The conceptual media-
tion of perception—Sense and reference and the concep-
tual and the empirical—Truth and practical adequacy—
Relativism, inter-theory disputes and discontinuities in
the development of knowledge—‘Theorizing’ and the
development of knowledge—Conclusions

3 Theory and method I: abstraction, structure and cause 85
Abstraction and structural analysis—Structure, agency
and reproduction—Contentless abstractions—Generali-
zation—Causation and causal analysis—Conclusions

4 Theory and method II: types of system and their
implications 118
Stratification and emergent powers—Closed and open
systems and regularities—Laws in science: causal and
instrumentalist—Prediction—Rational abstractions and
‘chaotic conceptions’—From abstract to concrete: the
example of marxist research—The theoretical and the

vii

viii Contents

empirical revisited—Spatial form and abstract and con-
crete research—Conclusion

5 Some influential misadventures in the philosophy
of science 153
Atomism and the problems of induction and causation—
Necessity—The accusation of ‘essentialism’—The limits
of logic—Popper and deductivism

6 Quantitative methods in social science 175
Quantification—Mathematics: an acausal language—
Accounting and quasi-causal models—‘Theoretical’ and
‘empirical’ models and closed and open systems—The
role of assumptions in models—Statistical methods—
Conclusions

7 Verification and falsification 204
Philosophical criticism—Existential hypotheses—Predic-
tive tests—Causal explanations and explanatory tests—
Interpretations—beyond evaluation?—Conclusions

8 Popper’s ‘falsificationism’ 226

9 Problems of explanation and the aims of social science 232
Explanation and the question of difficulty: I orthodox
conception—Research design: intensive and extensive
—Explanation and the question of difficulty: II critical
theory conception

Appendix: Notes on realism, writing and the future of
method in social science 258
Narrative versus analysis—The neglect of description—
The influence of rhetoric

Notes and references 267
Bibliography 299
Index 310

Preface to the second edition

In the 1980s, the ideas of realist philosophy began to make an
impact on social science. Yet the gulf between the more
philosophical debates and the literature on how we should do social
research remains wide, spanned by only the most rudimentary of
bridges. Sadly, many social scientists can still only think of ‘method’
in terms of quantitative techniques, and even though these are now
commonly supplemented by qualitative techniques such as
participant observation and informal interviewing, the basic
activity of conceptualization—which no one can escape—remains
unexamined. Of course realism has not had a monopoly of
innovations in philosophy and methodology in recent years.
Particularly important has been the growing interest in language,
writing and rhetoric, for these affect not merely how we re-present
ideas for others but the very terms in which we think.
Unfortunately these advances have been affected or infected by
idealist currents which appear to rule out the possibility of any kind
of empirical check on social science.
In view of this situation I believe that realism and the question of
method remain very much on the agenda and that there is still far
to go in developing a constructive discussion of method informed
by realist philosophy. This remains the task of this second edition.
The book is intended both for students and researchers familiar
with social science but having little or no previous experience of
philosophical and methodological discussions and for those who
are familiar with them but are interested in realism and method.
These two audiences have different interests and preferences
regarding style and content. The style and organization are
emphatically geared towards the first group (reviewers please
note!). I have therefore deliberately avoided spattering the text with

ix

x Preface to the second edition

name-droppings that would only alienate the first group even if
they reassured the second. Issues are selected on a need-to-know
basis rather than on one of fashion; philosophical doctrines are
only discussed if they have had or are likely to have a major
influence on the practice of social science. At the same time I feel
confident that the cognoscenti will find the realist ideas developed
here radically different from those dominant in the literature.
The two possible audiences are liable to ask different questions
and raise different objections. Those likely to come from the first
type of reader are anticipated and answered in the main text.
Answers to probable objections from the cognoscenti are restricted
to Notes and to Chapters 5 and 8, which provide critiques
specifically directed at certain orthodox ideas. The point of this
form of organization is to avoid the usual academic’s habit of
lapsing into writing only for specialists (including reviewers!). I
should also perhaps point out that although its arguments are often
philosophical, this book is primarily about method in social
research, rather than about the philosophy of social science. Many
fine books on the latter already exist.1 While they offer excellent
philosophical critiques they offer little constructive comment on the
practice of social science. It is this imbalance that I aim to redress.
A few words about revisions for those familiar with the first
edition. Second editions are an opportunity to update and another
chance to get things right and this is no exception. It’s common
today to acknowledge that texts and the way they are interpreted
can never be fully controlled by their authors, and often I have been
taken aback as much by supporters’ readings as by opponents’. But
authors do have some responsibility for the reception of their
books, so besides adding new material I have tried to correct my
own errors and to block some of the misreadings apparent in
reactions to the first edition.
The chief surprise to me about the reception of the first edition
has been the selectivity of interest. First, for reasons I still do not
fully understand, the necessary-contingent distinction introduced in
Chapter 3 seems to have overshadowed much of the rest of the
book. In this second edition I have tried to clarify this distinction
but I remain unconvinced that it warrants the prominence within
realism that some interpreters of the first edition gave it. The
second kind of selectivity involves a tendency to identify realism
with extraordinarily limited tendencies in social theory (e.g.
particular angles on marxism) and highly restricted areas of social

Preface to the second edition xi

research (e.g. research on localities). Whatever judgements were
made of this research—good or bad—seemed to have rubbed off
onto perceptions of realism. Let me therefore stress that, as any
scan of the literature will show, realism is a philosophy of and for
the whole of the natural and social sciences.
Reactions from students have made it clear that a new and fuller
Introduction was needed. Apart from this, the main additions
concern the nature of theory and its relation to empirical research,
practical knowledge, space and social theory, interpretive
understanding, research design and an appendix on realism and
writing. Further revisions have been made in the light of the
experience of empirical research carried out in the last six years.
Numerous minor changes have been made to correct and clarify
arguments, to add illustrations and to improve accessibility.

Acknowledgements

The University of Sussex for sabbatical leave; the University of
California, Los Angeles, Ohio State University, the universities of
Copenhagen, Roskilde and Lund and the Copenhagen Business
School, for their hospitality in providing me with new horizons; the
many graduate students in those places and the Sussex Concepts,
Methods and Values’ students for enduring my obsession with
methodology; and John Allen, Bjørn Asheim, Roy Bhaskar, Eric
Clark, Kevin Cox, Simon Duncan, Steen Folke, Frank Hansen,
Torsten Hägerstrand, Peter Maskell, Doreen Massey, Kevin
Morgan and Dick Walker, for their support, encouragement and
criticism. Finally, my love and thanks to Lizzie Sayer and Hazel
Ellerby.

Introduction

The status of social science is seriously in doubt. Outsiders’
attitudes towards it are often suspicious or even hostile, and social
scientists themselves are deeply divided over what constitutes a
proper approach to social research. The uncertainty has been
heightened by increasing doubts in philosophy about traditional
views of scientific objectivity and progress. Arguments about
whether social science should be like natural science no longer take
place on the basis of agreement about the nature and methods of
the latter. However, recent developments in realist philosophy have
offered new and productive perspectives in both areas that change
the whole basis of discussion. In this book I shall try to explain
these and show how they can resolve some of the problems that
have troubled social scientists.
One of the main difficulties of the existing literature on social
theory and the philosophy of the social sciences is that few
constructive contributions have been made on the subject of
method in empirical research, while texts on methods have
reciprocated this lack of interest by ignoring developments at the
philosophical level and in social theory. For example, much has
been written on theories of knowledge, but little about their
implications for empirical research. The result is that even where
the philosophical critiques have been accepted in principle they
have failed to make much difference in practice; indeed, the lack of
work on alternative methods has actually discouraged some of the
critics and their supporters from even venturing into empirical
research. Meanwhile, many of the empirical researchers whose
work has been under attack have been content to conclude that the
debate is not really relevant to them, or else that philosophical
discussions in general threaten empirical research and should

1

2 Method in social science

therefore be avoided. To get beyond this impasse we must decide
whether the critiques imply that we can continue to use the usual
empirical methods of hypothesis formation and testing, the search
for generalizations and so on, or whether these must be displaced or
supplemented by quite different ones. One of the chief aims of this
book is to answer these questions.
So much depends in social research on the initial definition of
our field of study and on how we conceptualize key objects.
Examples of these initial orientations include the adoption of lay
categories and classifications in sociology, the equilibrium
assumption in economics, the concept of the subject in psychology,
concepts like ‘interest group’ in politics, and the selection of spatial
units in human geography. All such starting points are fraught with
problems which, whether noticed or not, shape the course of
research long before ‘methods’ in the narrow sense of techniques
for getting and interpreting information are chosen. Once these
questions of conceptualization are settled—and frequently the
answers are matters of habit rather than reflection—then the range
of possible outcomes of research is often quite limited. These
matters are all the more difficult in social science where our
concepts are often about other concepts—those of the society that
we study.
In view of this it is quite extraordinary to compare the attention
given in social science courses to ‘methods’ in the narrow sense of
statistical techniques, interviewing and survey methods and the
like, with the blithe disregard of questions of how we
conceptualize, theorize and abstract. (‘Never mind the concepts,
look at the techniques’ might be the slogan.) Perhaps some would
be content to dismiss these matters as questions of paradigms,
social theory or intuition, not method, but it is my belief that there
is method not only in empirical research but in theorizing, and that
we need to reflect on it.
A second major impediment to the development of effective
method in social science concerns causation. So much that has been
written on methods of explanation assumes that causation is a
matter of regularities in relationships between events, and that
without models of regularities we are left with allegedly inferior, ‘ad
hoc’ narratives. But social science has been singularly unsuccessful
in discovering law-like regularities. One of the main achievements
of recent realist philosophy has been to show that this is an
inevitable consequence of an erroneous view of causation. Realism

Introduction 3

replaces the regularity model with one in which objects and social
relations have causal powers which may or may not produce
regularities, and which can be explained independently of them. In
view of this, less weight is put on quantitative methods for
discovering and assessing regularities and more on methods of
establishing the qualitative nature of social objects and relations on
which causal mechanisms depend. And this in turn, brings us back
to the vital task of conceptualization.
Social scientists are invariably confronted with situations in
which many things are going on at once and they lack the
possibility, open to many natural scientists, of isolating out
particular processes in experiments. Take an apparently simple
social event such as a seminar. It involves far more than a discussion
of some issues by a group of people: there is usually an economic
relationship (the tutor is earning a living); students are also there to
get a degree; their educational institution gets reproduced through
the enactment of such events; relations of status, gender, age and
perhaps race are confirmed or challenged in the way people talk,
interrupt and defer to one another; and the participants are usually
also engaged in ‘self-presentation’, trying to win respect or at least
not to look stupid in the eyes of others. This multi-dimensionality
is fairly typical of the objects of social science. The task of assessing
the nature of each of the constituent processes without being able
to isolate them experimentally throws a huge burden onto
abstraction—the activity of identifying particular constituents and
their effects. Though largely ignored or taken for granted in most
texts on method I believe it to be central.
I shall therefore take a broad view of ‘method’ which covers the
clarification of modes of explanation and understanding, the nature
of abstraction, as well as the familiar subjects of research design
and methods of analysis. The terrain of the discussion is therefore
the overlap between method, social theory and philosophy of social
science.
In view of this overlap many of the arguments have a
philosophical character, involving thinking about thinking. But
while I believe social scientists can learn from philosophy they
should not be in awe of it, for they can also inform it. (Much
damage has been done by prescriptions made by philosophers who
have little or no knowledge of what social science involves.)
Methodologists need to remember that although method implies
guidance, research methods are the medium and outcome of


research practice;1 the educators themselves have to be educated—
with frequent refresher courses. Therefore philosophy and
methodology do not stand above the substantive sciences but serve,
as the realist philosopher Roy Bhaskar put it, as ‘underlabourer and
occasional midwife’ to them.2 And social scientists should certainly
not fear that philosophical thinking will subvert empirical research,
though it may be heavily critical of certain kinds.
Method is also a practical matter. Methods must be appropriate
to the nature of the object we study and the purpose and
expectations of our inquiry, though the relationships between them
are sometimes slack rather than tight. If we imagine a triangle
whose corners are method, object and purpose, each corner needs
to be considered in relation to the other two. For example, what do
differences between the objects studied by social and natural
sciences imply for the methods they use and the expectations we
have of their results? Is the goal of prediction appropriate to an
object such as an ideology? Can social scientific method ignore the
understandings of those whom it studies? How far would an
interpretive, ethnographic method be appropriate for assessing
macro-economic change? To answer such questions we shall have
to consider all three corners of the triangle.
Although methodology needs to be critical and not merely
descriptive I intend to counter various forms of methodological
imperialism. The most important kind, ‘scientism’, uses an absurdly
restrictive view of science, usually centring around the search for
regularities and hypothesis testing, to derogate or disqualify
practices such as ethnography, historical narrative or explorative
research, for which there are often no superior alternatives.
Another kind of imperialism, formed in reaction to this is that
which tries to reduce social science wholly to the interpretation of
meaning. A critical methodology should not restrict social science
to a narrow path that is only appropriate to a minority of studies.
The variety of possible objects of study in social science stretches
beyond the scope of a single model of research. Consequently, while
this book is about method it is not a recipe book, though it is
intended to influence the construction of recipes for research, by
suggesting ways of thinking about problems of theorizing and
empirical research. Examples are therefore intended as just that—
not as unique restrictive moulds to which all realist research must
conform.
But what is realism? First of all it is a philosophy not a

Introduction 5

substantive social theory like that of Weber or neoclassical
economics. It may resonate more with some social theories than
others (e.g. marxism more than neoclassical economics) but it
cannot under-write those with which it appears to be in harmony.
Substantive questions like ‘what causes inflation?’ are different
from philosophical questions like ‘what is the nature of
explanation?’
Things get more difficult when we try to define the content of
realism. When confronted with a new philosophical position for the
first time it is impossible to grasp much of what is distinctive and
significant about it from a few terse statements of its
characteristics. Particular philosophies are not simple and self-
contained but exist through their opposition to a range of
alternative positions. They involve loose bundles of arguments
weaving tortuously across wider fields of philosophical discourse.
Nevertheless, readers may prefer to have at least some signposts
regarding the nature of realism, or rather my own view of it, even
if their meaning is limited at this stage. Some of the following
characteristic claims of realism may seem too obvious to be worth
mentioning, but are included because they are in opposition to
important rival philosophies. Some may seem obscure, but they
provide at least some orientation to newcomers to realism. Fuller
explanations will come later. The wordings represent a compromise
between what would be acceptable to those familiar with
philosophical discourse and what is likely to be accessible to those
new to it.

1 The world exists independently of our knowledge of it.
2 Our knowledge of that world is fallible and theory-laden.
Concepts of truth and falsity fail to provide a coherent view of
the relationship between knowledge and its object. Never-
theless knowledge is not immune to empirical check, and its
effectiveness in informing and explaining successful material
practice is not mere accident.
3 Knowledge develops neither wholly continuously, as the steady
accumulation of facts within a stable conceptual framework,
nor wholly discontinuously, through simultaneous and
universal changes in concepts.
4 There is necessity in the world; objects—whether natural or
social—necessarily have particular causal powers or ways of
acting and particular susceptibilities.


5 The world is differentiated and stratified, consisting not only of
events, but objects, including structures, which have powers
and liabilities capable of generating events. These structures
may be present even where, as in the social world and much of
the natural world, they do not generate regular patterns of
events.
6 Social phenomena such as actions, texts and institutions are
concept-dependent. We therefore have not only to explain their
production and material effects but to understand, read or
interpret what they mean. Although they have to be interpreted
by starting from the researcher’s own frames of meaning, by
and large they exist regardless of researchers’ interpretations of
them. A qualified version of 1 therefore still applies to the
social world. In view of 4–6, the methods of social science and
natural science have both differences and similarities.3
7 Science or the production of any other kind of knowledge is a
social practice. For better or worse (not just worse) the
conditions and social relations of the production of knowledge
influence its content. Knowledge is also largely—though not
exclusively—linguistic, and the nature of language and the way
we communicate are not incidental to what is known and
communicated. Awareness of these relationships is vital in
evaluating knowledge.
8 Social science must be critical of its object. In order to be able
to explain and understand social phenomena we have to
evaluate them critically.

Amplifications of these points could fill many books but the list
should provide some orientation.
No book of this kind can expect to be exhaustive in its coverage
of the range of methodological issues of interest to social science or
of the types of social research to which they might be relevant. As
regards the latter, it is quite extraordinary how sociology has had
the lion’s share of attention in the literature. (Some authors give the
impression that social science is reducible to sociology and
sociology to the work of Durkheim, Weber and Marx!) This has
produced a deafening silence on the social research practice of
those in other disciplines such as economics, development studies,
psychology and human geography. While I cannot address all of
these I shall try to counter the usual sociological imperialism found
in most books on method in social science.

Introduction 7

Any author in this field works with implicit exemplars of
particular areas of social research. Mine are somewhat different
from those of existing texts; they come mostly from political
economic theory and interdisciplinary studies of industry and urban
and regional systems, in which researchers tend to come from
geography, sociology, economics, political science and
anthropology. However, no special knowledge of these is needed to
understand the examples I have used and indeed many of them
come from everyday arguments and events. I have deliberately
avoided the philosopher’s irritating habit of using trivial examples
(‘the tree in the’quad’, etc.). If a philosophical point is worth
making it may as well be illustrated by an example which not only
gives clarification but suggests its social and practical significance.
A few words are needed on terminology. At the centre of social
science’s internal crisis have been attacks on orthodox conceptions
usually termed ‘positivist’ or ‘empiricist’. So many different
doctrines and practices have been identified with these terms that
they have become devalued and highly ambiguous, or even purely
pejorative. Those who want to continue using them increasingly
find that they have to preface arguments with tiresome digressions
on ‘the real meaning of positivism’ and these often generate more
heat than what follows. I have therefore avoided using these terms
for the most part. This need not prevent one from discussing some
of the issues covered by them and indeed it is liberating to avoid the
usual burden of unwanted associations that the terms bear. In
general I have minimized the use of technical terminology. (That’s
what they all say, I know, but at least the intention was there!)
The word ‘science’ needs special comment. There is little
agreement on what kinds of methods characterize science beyond
the rather bland point that it is empirical, systematic, rigorous and
self-critical, and that disciplines such as physics and chemistry are
exemplars of it. Most users of the term obviously consider it to
have strong honorific associations for few are willing to cede its use
to opponents. Those who want to stand apart from the futile
academic game of trying to appropriate and monopolize this
descriptively vague but prized label for their own favoured
approaches are liable to be accused of the heresy of not caring
about science and, by implication, rigour and other virtues. While
no one is likely to be against virtue, the coupling with exemplars
like physics is particularly unhelppful. Not only is there little
consensus on what their methods are, it is also not self-evident that


they are appropriate for the study of society; indeed, that very
question has been at the heart of the philosophical debates. The use
of the word ‘science’ in this strong sense has allowed many authors
to prejudge precisely what has to be argued. I therefore want to
make it clear that ‘science’, ‘natural science’ and ‘social science’ are
used in this book simply as synonyms for the disciplines that study
nature and society. At the most, these subjects might be said to
distinguish themselves from everyday knowledge by their self-
examined and inquisitive character; but that does not say very
much and proponents of the humanities may want to include
themselves in this description. In other words, my lack of
commitment in the use of the word ‘science’ does not, of course,
entail any lack of commitment to the search for rigorous and
effective methods of study; rather it is intended to clear away an
important obstacle to their discovery.
In view of my attacks on the insulation of discussions of method
from social theory and philosophy of science, readers will not
expect me to plunge immediately into a discussion of particular
methods or techniques. In Chapter 1 we look at knowledge in
context, situating social scientific knowledge in relation to other
kinds and to practice. Any theory of knowledge is handicapped
from the start if it ignores this context for it is likely to ignore how
the internal structure and practices of science are shaped by this
position. And it is a particularly important consideration for studies
of society, for everyday knowledge is both part of their object and
a rival source of explanations. A discussion of the nature of the
relation between subject and object in social and natural science
then provides a basis for an introduction to the necessarily
interpretive and critical character of social science.
Having looked at the context of knowledge, Chapter 2 examines
some dominant views of its status and reliability. The time when
science was thought to involve the steady accumulation of objective
knowledge through a neutral medium of observation has long since
gone. In its place there has been a crisis of confidence in which
relativism and doubts about the possibility of empirical evaluation
and scientific progress have been rife. We begin from the point at
which most popular discussions confront the problem -the nature
of facts, observation and theory and the relationship between them.
To make any progress on this, and in order to say anything sensible
about method, particular attention has to be paid to the meaning of
‘theory’ (woefully underexamined in the philosophical and

Introduction 9

methodological literature), and to the linguistic and practical
character of knowledge. Traditionally doubts about objectivity and
the status of scientific knowledge have involved arguments about
the nature of truth and how it might be established. In our case we
shall approach these matters differently, attempting to counter the
neglect of the linguistic and practical character of knowledge,
arguing that the concept of truth (and falsity) is incoherent, and
that knowledge needs to be evaluated in terms of ‘practical
adequacy’. The chapter ends with an assessment of the problem of
relativism and the resolution of inter-theory disputes.
This prepares the ground for a more focused discussion of
method in the ensuring chapters. In these we move continually
between the three points of our triangle of method, nature of the
object and purpose of study. Following our emphasis on the activity
of conceptualization and theorizing we begin in Chapter 3 at the
most ‘primitive’ level with an important but under-analysed aspect
of it—abstraction and the relation between abstract and concrete
research. We then consider the nature of social relations and
structures and how abstraction can illuminate them. We then
clarify the nature of generalization, with which abstraction is
commonly confused. The chapter ends with a discussion of the
realist concept of causation in social science and its implications for
methods of causal analysis.
Chapter 4 considers method in relation to ontology or the nature
and structure of the social and natural world: first, in so far as it is
‘stratified’ so that certain objects, such as institutions, have powers
emergent from, or irreducible to, their constituents; second, in so
far as it consists of ‘open systems’ in which regularities in events are
at best approximate and transitory. The implications of these
characteristics for the possibility of discovering laws and for
explanation and prediction in social science are then assessed.
Further implications of ontological matters for method are then
examined: ‘rational abstraction’ and the need to make abstractions
sensitive to the structure of their objects; the relationship of theory
and empirical research to the discovery of necessity in the world;
and the consequences and dangers of the abstraction from space
and time in social science.
Chapter 5 is a digression from the main argument of the book.
It is included for those readers who are familiar with more
orthodox positions in philosophy and methodology and who may
require answers to certain objections which these raise before


proceeding any further. Others may wish to ‘fast forward’ to
Chapter 6. The main issues concern a connected set of problems in
mainstream philosophy of science, many of them particularly
associated with the work of Karl Popper, who has been particularly
influential in social science: induction, atomistic ontology,
causation, necessity, essentialism, logic and deductivism.
In Chapter 6 we turn to quantitative methods. As before, and in
contrast to the usual treatment in texts on method, these are
evaluated in relation to their appropriateness to the nature of the
object of study, the scope for quantification and the implications of
open systems for modelling. The discussion then opens out into a
critical assessment of the use of models themselves and the role of
assumptions. Lastly I examine the resonances between the use of
quantitative positions and particular views of society as atomistic
and views of method which misguidedly focus on the search for
regularity and neglect conceptualization and interpretive
understanding.
The evaluation, or verification and falsification, of social
scientific accounts and theories is the subject of Chapter 7. In
accordance with our emphasis on the diversity of appropriate
methods, we argue that evaluation is a complex and differentiated
business, varying according to different objects of study and types
of claim. Chapter 8 is a second digression for readers familiar with
orthodox philosophy of science, presenting a critique of Popperian
views of falsification.
In Chapter 9, we return to problems of explanation in social
science. Explanations are shown to be characteristically incomplete
and approximate and to vary according to the relationships of our
triangle of method, object of study and purpose of research. Yet
researchers often over-extend particular approaches, for example in
expecting too much of generalization. I therefore discuss the limits
and interrelations between key types of research, and try to
illuminate them by comparing the capabilities of different kinds of
research design. The chapter concludes by returning to the wider
context of knowledge with which we began: ultimately our
judgements about problems of explanation depend in part on
whether we accept or try to resist the critical and emancipatory role
of social science.
Finally, in the Appendix, I comment on some implications of
recent interest in the fact that scientific knowledge is usually
presented in the form of texts. Arguably, the rhetoric we use and the

Introduction 11

form in which we present knowledge are not neutral carriers of
meaning but influence the content. Ways in which this can happen
are illustrated briefly. Contrary to many commentators, I argue that
while these concerns do indeed require further attention, they need
not threaten realism.

1 Knowledge in context

We feel that even when all possible scientific questions have been
answered, the problems of life remain completely untouched.
(Wittgenstein, 1922, 6.52)1
‘Method’ suggests a carefully considered way of approaching the
world so that we may understand it better. To make judgements
about method it helps considerably if we have some idea of the
nature of the relationship between ourselves and that which we
seek to understand. Yet it is at this fundamental level that many
arguments about method go wrong, for they fail to consider
knowledge in its context.
How does social science relate to everyday knowledge in society
and to natural science? Does it merely mystify or reproduce the
former? Should it emulate the latter? Some of those who have
attacked social science for the alleged triviality of its findings and
for lacking relevance to practical matters have argued that this is
due to its failure to use the ‘proven’ methods of natural science.
Others have argued that triviality is precisely the result of using
such methods. There is disagreement about whether it should adopt
a ‘disinterested’ stance with respect to practice or be actively
involved in the process of social development. Some see social
science as a natural science of society which can be applied through
social engineering. Others see their role as having more in common
with a therapist than an engineer, their aim being the development
of greater self-understanding. Still others consider the role of social
science to be the critique of society.
In this chapter, I shall examine in abstract terms2 the context in
which knowledge, especially social science, develops and how it
relates to practice and to its objects. This, I hope, will provide a
basis upon which the above problems can be discussed in this and

12

Knowledge in context 13

later chapters. Some of the questions posed here might seem
strangely broad, even for philosophical discussions, and
superficially some of the answers may appear obvious. But if such
points are ignored or taken for granted, we may fail to notice how
they challenge some of the underlying assumptions of social
science’s practice. Indeed, their significance goes beyond academia
to everyday life, for they suggest that in certain ways society
systematically misunderstands itself.
One of the most extraordinary features of the literature on the
methodology and philosophy of science is the extent to which it
ignores practice and the way in which knowledge is involved in
what scientists and lay people do. If, as is the custom of this
literature, we reduce practice to knowledge, knowledge to science,
and science to observation and contemplation, then it is small
wonder that it should prove difficult to assess the relation between
the social and natural sciences and their objects. Although there is
far to go in working out the implications of the practical context of
knowledge, I wish at least to set out on this road.3

Some misconceptions about knowledge

I shall start by combating the following (interrelated)
misconceptions:

1 that knowledge is gained purely through contemplation or
observation of the world;
2 that what we know can be reduced to what we can say;
3 that knowledge can be safely regarded as a thing or product,
which can be evaluated independently of any consideration of
its production and use in social activity;
4 that science can simply be assumed to be the highest form of
knowledge and that other types are dispensable or displaceable
by science.

1 and 2 are highly interrelated and together constitute the
‘intellectualist fallacy’ or ‘prejudice’. All four misconceptions help
to make the relationship between social science and society
problematic.
Against 1, I shall argue that knowledge is primarily gained
through activity both in attempting to change our environment


(through labour or work) and through interaction with other
people, using shared resources, in particular a common language.4
Although the development of knowledge may be furthered through
passive contemplation of the world, it always presupposes the
existence of these two contexts, which provide a kind of feedback
or test for our ideas and a language in which and with which to
think. Individuals cannot develop knowledge independently of a
society in which they can learn to think and act. The nearest
approximation to the unsocialized individual in human experience
is the ‘wolf-child’ who, having largely been brought up outside
human society, is often scarcely able to walk on two legs, let alone
speak or perform the simplest tasks of reasoning.
In so far as people and their ideas are included among our
objects of knowledge, the relationship of knowledge to practice
may be interactive rather than passive and purely reflective. It is
particularly clear with self-reflection that in thinking about
ourselves, we can change our ‘object’. Under certain conditions,
social science can have a similar effect on its object. Moreover, the
search for truth, the attempt to rid social knowledge of illusion,
puts reflective, examined knowledge into a critical relationship
with false beliefs and their effects in society. In this sense the role of
social science and perhaps also the humanities may be critical,
therapeutic and even emancipatory. For example, arguments about
the meaning of masculinity and femininity, about the nature of
economic recession or about international politics don’t take place
outside society as competing external descriptions: they are part of
the social process itself. I will develop these points shortly.
Another aspect of the contemplative view of knowledge is the
assumption that the only function of knowledge and language is
‘prepositional’ 5 (to make propositions about the world) or
‘referential’. What is overlooked in this view is that knowledge
concerns not only ‘what is the case’ or ‘knowing-that’ but ‘know-
how’, that is knowing how to do something, whether it be physical
behaviour or communicating successfully with others.
Misconception 2, the second component of the intellectualist
fallacy, follows this closely. It concerns the tendency to pedestal
spoken or written forms of knowledge and to imagine that these are
the only ways in which meaning can be communicated and
knowledge can be ‘carried’ and applied. With this goes a tendency
to derogate those types of practical knowledge which do not
require much linguistic competence, but which nevertheless involve


practical skills. Much of everyday knowledge takes this practical
form: a young child learns a great deal before it acquires a
language; we have many skills which we are aware of and yet
cannot describe verbally and also many of which we are usually
unaware. Not all social behaviour is acquired and mediated
linguistically, even in the form of talk internalized in our heads.
Much of what we do does not proceed on the basis of a model of
‘rational choice’but involves a learned accommodation to familiar
circumstances which, as Bourdieu puts it,

[is]. . .neither the outcome of the explicit aiming at consciously pursued
goals, nor the result of some mechanical determination by external causes
. . .[but]. . .guided by a practical sense, by what we may call a feel for the
game.6

Social scientific knowledge is primarily prepositional or referential,
rather than practical, and this should immediately provide some
clues as to why it seems unable, except very indirectly, to help us
decide how to live. No doubt the common fear of the alleged
danger of ‘value intrusion’ in social science also inhibits its practical
application.
There are also material circumstances which reinforce this
intellectualist prejudice. Academics generally occupy a place in the
social division of labour in which the development of knowledge in
prepositional forms, in a contemplative relationship to the world,
has unusual primacy. Within this restricted but privileged context,
the activities of speaking and writing are elevated above those of
making and doing, as if it were possible to live on prepositional
knowledge and linguistic communication alone. Not surprisingly,
as we shall see, social scientists, philosophers or intellectuals
frequently project these characteristics onto society as their object
of study, underestimating the extent to which social behaviour is
guided by a vague and unexamined practical consciousness.7 Social
scientists may examine it but the results of that examination should
not be confused with the original and projected back onto it, or
divorced from its practical setting. We shall have more to say about
these problems in Chapter 3. Despite the extent of the freedom of
academics to reflect upon almost anything, the restricted horizons
of their place in the social division of labour encourage a blind spot
where practical and tacit skills are concerned. The slanting of our
educational system towards a one-sided emphasis of an


intellectualist and linguistic view of intelligence and skill is partly
attributable to this.
Having written this, in a book I can obviously only combat this
prejudice from within!
Misconception 3 concerns the common tendency to think of
knowledge as a product or thing which exists outside of us, which
we can ‘possess’ and which is stored in finished form in our heads
or in libraries. We tend not to think in terms of knowing, which is
in the process of becoming, ‘in solution’, as consciousness, but as a
thing already ‘precipitated’. 8 Despite the work involved in
developing and sharing knowledge, this active side (perhaps again
as a result of the intellectualist prejudice) tends to be overlooked.
As such, it is an instance of the common tendency to reify the social
world; that is, to turn active, conscious social relationships and
processes into things which exist independently of us so that we
think of them in terms of ‘having’ rather than ‘being’.9 Although,
for the sake of accessibility, I have used the reified noun-form
‘knowledge’ in preference to the unreified but unfamiliar and
ambiguous ‘knowing’, I shall try to counteract the misconceptions
which it can encourage.
To combat this static view it is imperative to consider the
production of knowledge as a social activity. 10 To develop
‘knowledge’ we need raw materials and tools on which and with
which we can work.11 These are linguistic, conceptual and cultural
as well as material. In trying to understand the world, we use
existing knowledge and skills, drawn from whatever cultural
resources are available, to work upon other ‘raw’ materials—
knowledge in the form of data, pre-existing arguments, information
or whatever. It is only by this activity, this process, that knowledge
is reproduced or transformed: it is never created out of nothing. To
paraphrase Bhaskar, knowledge as a product, a resource, a skill, in
all its various forms, is ‘both the ever-present condition and
continually reproduced outcome of human agency’.12 Science is not
a thing but a social activity.
The fourth common misconception about knowledge concerns
scientism.13 Despite the fact that philosophy is generally taken to
allow no limitations on what it can question, there is a striking
tendency in Anglo-American philosophy of science and social
science simply to assume that science is the highest form of
knowledge, to which all should aspire. Again, this resonates with
and reinforces the intellectualist prejudice. A large number of texts


on the philosophy of science take this as their point of departure
and immediately pass on to the description or prescription of its
internal procedures. But this unquestioning attitude towards the
status of science and how it relates to other kinds of knowledge can
prejudice the whole discussion of the internal questions of
procedures of empirical study, modes of inference, models of
explanation and testing etc.
I shall argue that different types of knowledge are appropriate to
different functions and contexts; for example, engineering for the
task of making nature move to our designs, ethics to the
harmonization of the conduct of people in society. But these
contexts are not mutually exclusive but overlapping. Scientific
practice embraces several types of knowledge, including some
which are generally excluded as non-science or even anti-science by
scientism. For example, many philosophers who have adopted this
stance of ‘scientism’ have treated ethical decisions as a-rational,
purely emotive and not part of science, which by contrast deals
purely with matters of fact, with rational and objective questions of
‘what is the case’. Yet science is also a specialized type of social
activity and as such it requires rules governing what is proper and
improper conduct; without ethical principles such as those
concerning honesty of reporting and refusal of illogical argument,
science could not exist. In other words, scientific knowledge
presupposes among its very foundations a kind of knowledge which
‘scientism’ has sought to deny, exclude or derogate.14 We will return
to other excluded but overlapping forms of knowledge shortly.
Having discussed some of the different kinds of knowledge, let
us now look at the context in which it develops and see what effect
it has.

Knowledge, work and communicative interaction

Knowledge is developed and used in two main types of context—
work (or ‘labour’) and communicative interaction.15 These contexts
are highly related but neither is wholly reducible to the other. By
‘work’ or ‘labour’, I mean any kind of human activity which is
intended to transform, modify, move or manipulate any part of
nature, whether it be virgin nature or nature that has already been
extensively modified; that is, whether it be mining, transport,
making and using machines, or putting letters in envelopes. All of


these activities involve the manipulation of matter for human
purposes.
Human labour, unlike the behaviour of animals, is conscious; the
worker has some conception of the goal, the end product of the
labour.16 Even where the labour has become thoroughly habitual,
this goal can be recovered. We can not only monitor the progress of
our material works; we can record and reflect upon our
monitorings, discuss them with others and generate new methods,
goals or projects to work on. The process of ‘knowing’ in this
context derives a certain kind of check through feedback from the
results of the work—not just through observing the world passively
as if it were external to us, in order to see if our knowledge
‘mirrors’ it successfully—but from the results of material activity as
one of nature’s forces, operating within nature. Natural science
itself is by no means just a matter of observation and
conceptualization; its practitioners spend most of their time
intervening in nature, doing things to it, trying to make
experiments work.17 In monitoring and checking the practical
knowledge that we use in work, what is at issue is the success or
failure of this transformation—this active ‘objectification’ of
knowledge—rather than a passive ‘mirroring’ or ‘representation’ of
the world. This, in turn, should affect how we evaluate or test
knowledge: The question whether objective truth can be attributed
to human thinking is not a question of theory but a practical
question. In practice man must prove the truth, i.e. the reality and
power…of his thinking.’18
Given that human life depends on it, work, as the
transformation of nature for human purposes, gets surprisingly
little attention in philosophy and even in social science. This might
be an instance of the academics’ projection of their own way of life
on to the lives of those they study. It is not only films and popular
fiction that tend to neglect the means by which people earn their
living. Many social theories pay great attention to how society is
organized and how it coheres, without considering how people
(re)produce their means of life. Yet work is the most transformative
relationship between people and nature. It is both a material
process and a conscious one: it cannot be reduced either to pure
physical behaviour or passive contemplation.19 It is a ‘missing link’
that bridges the gap between knowledge and the world—a gap
which has been widened both by the intellectualist prejudice and
the real separations of work and ‘living’ of capitalism.


Labour is also central to an understanding of human
development or ‘self-change’. In changing our social and natural
milieux we change the forces and conditions which shape the
character of society and its people. As new kinds of work and social
relations develop, people develop new needs. In other words,
human beings have a capacity for ‘self change’, for making their
own history, though as Marx noted: ‘they do not make it just as
they please, they do not make it under circumstances chosen by
themselves, but under circumstances directly encountered, given
and transmitted from the past’.20 In other words, history not only
happens to people but is made by them, consciously or
unconsciously. Any conception of society—whether lay or
scientific—which treats people as passive objects of history and
mere carriers of knowledge, rather than agents or producers, is
doomed to misrepresent both its object and itself.
The second basic context of knowledge is ‘communicative
interaction’. By this I mean any kind of interaction between people
which involves the sharing or transmission of meaning. It is by no
means limited to spoken or written communications, but includes
many kinds of activity which presuppose understanding the
meaning of signs, conventions, concepts, pictures, rules and
actions. Even where the communication is linguistic, there is often
an important non-verbal dimension. An obvious example is in job
interviews, where both interviewer and interviewee draw upon a
wide range of social skills of interpretation, self-presentation and
‘impression-management’ 21 in addition to those involved in
speaking.
Paradoxically, while it has been common to ignore knowledge
which is not expressed in language, until recently social scientists
and methodologists have taken the linguistic character of their own
knowledge for granted, as if language were nothing more than a
transparent and unproblematic medium. On reflection it seems
extraordinary that methodology should treat the ability to use
language effectively as irrelevant to our ability to understand and
explain the world. The attention normally given to technical
methods of analysis is in gross disproportion to the consideration
given to the language in which we characterize the world.
Language therefore needs to be put in its place, elevated from its
present position of neglect, though not abstracted from its
context.22
First of all, language has effects of its own, which go beyond


those intended by users. The possible meanings that spring from the
interaction between the play of associations among the various
components of language and contexts depend in part upon the
structure of language. We are accustomed to thinking of language
as something which we, as users, speak with and through. But there
is a sense in which the reverse applies too; I am not the sole author
of this book: the structure of language and narrative forms, such as
those of academic texts, of which I am only partially aware, speak
through me. At one level we might say that this is analogous to any
act of production, such as the construction of a house, for the
nature of the materials, as well as the work of the builder,
determine the properties of the result. But the effects of language
are not fixed like those of bricks and steel. New interpretations are
always possible; they can never be foreclosed.
Secondly, language cannot exist for an isolated individual who
has never been socialized, for language is both a medium and
product of social interaction. 23 Propositional knowledge is
constructed and expressed in terms of the concepts available in a
language and we seek inter subjective confirmation of the
propositions through communicative interaction. In scientific
communities this kind of checking is highly formalized in order to
strive for rigour of thought.
Thirdly, language also has an expressive function. Although the
expression of feelings may seem particularly personal or individual,
it is nevertheless done in the terms available in one’s language and
hence has a social dimension.
Fourthly, much of our knowledge and our uses of language
concern neither making propositions about the world nor
expressing our feelings but rather have a directly social function
through providing the means by which we question, command,
argue, confer respect or distribute contempt, establish relationships
and generally conduct our business in society.24 In no case can
knowledge or language be treated as if they existed outside the
social context. Even if our interest (like many philosophers’) is
primarily in the truth or falsity of knowledge ‘regardless of its
social origins’ it must be remembered that judgements of truth or
falsity require intersubjective appraisal.
For analytical and expositional convenience, I have dealt with
these two contexts of knowledge of labour and communicative
interaction separately. This gives us only a very provisional, crude
outline, for the two are in fact interdependent. The development of


human labour from merely animal behaviour requires the
simultaneous development of a high level of communicative
interaction through which people can acquire and develop the
‘instrumental’ knowledge which they use in labour.
Systems of meaning are negotiated by people in the course of
social interaction.25 As such these systems have a conventional
character—they become conventions according to which actions of
individuals can be related; the systems of meaning related to money
are a good example. However, not just any conventions will do;
those which can inform successful labour and interaction which we
need to undertake to survive will be preferred, while those which (it
is intersubjectively agreed) cannot inform successful projects will be
winnowed out. It is because nature and its material processes
(including human activity) have particular structures and properties
which exist independently of our understanding of them, that not
just any understanding will serve as a basis for activity.26 Through
intersubjectively monitoring our interventions in nature we try to
develop our language and knowledge in accordance with those
activities which seem practically possible. The presence of power
and domination in the social determination of meaning modifies
this situation only slightly, for the powerful are bounded by the
realm of the possible too. I will return to and develop these points
more fully later.
Although human labour and communicative interaction are
highly interdependent, we cannot collapse one into the other.27 At
the limit, even though communication can be hard work (!), it
cannot be reduced wholly to the material transformation of the
world. Even though the interpretation of meaning and the most
passive forms of contemplation involve material processes in the
brain, meaning is not reducible to them. Even if you could observe
the chemical and physical processes at work in someone’s brain as
they spoke, you would still need to know the meaning of what they
said in order to be able to understand them. Conversely, work as
the transformation of matter cannot be wholly reduced to the
sharing or interpretation of meaning.
Once again, misconceptions about the context of knowledge can
distort social scientists’ views of both their object of study and their
own activity. An approach called Radical behaviourism’ provides a
good example: its proponents insist that the meanings people attach
to their actions and to other objects play no part in determining
what they do. Knowledge is therefore divorced from practice. This,


of course, raises the question of the radical behaviourists’ view of
their own activity—have their ideas nothing to do with their
actions? This is an extreme case whose absurdity is clear enough,
and usually the misconceptions are less obvious. Nevertheless, it is
certainly not unusual for social scientists to ignore many of the
meanings people attach to situations, although few would insist on
doing so as a matter of principle. In discussions of philosophy and
methodology few accept radical behaviourism, but in actual social
scientific practice something approaching it is common,
particularly in the work of those who see their task as the search for
law-like empirical regularities equivalent to those found in some of
the natural sciences. It is therefore important to explore the
misconceptions further.

The relationship between subject and object

This account of ‘knowledge in context’ can be developed and
further clarified by examining the relationship between ‘subject’
and ‘object’. In most discussions of this, the term ‘subject’ (or
sometimes ‘knowing-subject’) refers to the observer or investigator
or simply ‘thinker’, while the ‘object’ is defined as the thing being
studied. I want to make two qualifications or additions to these
definitions. First, as before, I do not want to restrict the meaning of
‘subjects’ to scientists, on the grounds that I want to bring out
similarities and connections between scientific and other kinds of
knowledge at this stage. Second, I want to include the older
meaning of ‘subject’, as a creative agent who brings about change.
The point of this modification is to avoid restricting the conception
of the relationship to a passive, contemplative mode from the start.
I will begin by introducing and criticizing some naïve
conceptions of the relationship and then go on to develop
alternative conceptions as they apply to natural and social science.
This will lead into a discussion of the differences and similarities of
natural and social science and of the contrasting approaches to
them, and finally bring us back to the problem of how social
science relates to everyday knowledge and practice.
Behind most views on this topic lies a conceptual framework
which includes the following series of dualisms or dichotomies:


people — nature
individual — society
subjective — objective
thought — action
mental — material
mind — body
knowledge — practice
beliefs — facts
expressive function — referential/propositional
of language function of language

This framework of oppositions is deeply embedded in our culture;
indeed it is difficult to think outside it. It is not only implicit in
common-sense thinking but explicit in much of British and
American literature on philosophy and social science. Never-
theless, although these dualisms are ‘second nature’ to us and
probably look quite harmless, I shall argue that every one of them
is beset with misconceptions which generate problems in our
understanding of the world and of ourselves. The dualisms do not
operate singly but in parallel, providing mutual reinforcement, so
that in the vertical dimension of the diagram, meanings or
associations ‘leak’ from one term to the next.
I have already alluded to some of the problems generated by this
framework, but I have hardly begun to draw out the implications.
These include the following:

1 Work and activity are excluded and banished to a kind of
limbo, so that people are separated from society and their own
activity, making it difficult for us to understand how thought
actually relates to and functions in nature and society. This
implies not only an inadequate theory of knowledge
(epistemology) but an alienated view of ourselves.
2 The framework is also alienating because the exclusion of
social relations and mtersubjectivity tends to reduce society to
nothing more than a group or loose aggregate of individuals.
At the same time it obscures the social function of language.
Indeed, the omission of intersubjectivity, as the context in
which language is (re)produced, makes language in general
difficult to comprehend.


These points can be substantiated in the course of a critique of
models of the subject-object relationship.
The simplest model fits comfortably within this conceptual
framework (Figure 1), where 3, the subject, observes and records
information about O, the object. On the basis of our earlier
arguments we can amend this so that the relationship includes
activity, particularly labour.

Figure 1 Subject and object: 1

It was also argued that the subject must have a language in
which to think about the object.28 Given the social nature of
language, the subject-object relationship in Figure 1 must
presuppose the existence of social relations, or ‘subject-subject
relations’29 within some language community. Usually the language
community is internally differentiated, embracing specialist sub-
groups with some of their own linguistic and conceptual resources,
be they those of physics, economics, farming, cooking, computer
programing or whatever. As this social context is not incidental but
indispensable to the subject-object relationship, we shall modify the
diagram accordingly, assuming for the time being, for the sake of
simplicity, that O consists only of non-social objects (Figure 2).
Figure 2 echoes the points made earlier about work and
communicative interaction as interdependent contexts of
knowledge, for it shows that subjects (whether laypersons,
specialists, academics or whatever) stand in a double relationship—
to their object and to other subjects. Subjects cannot gain
prepositional knowledge of their objects or acquire practical
knowledge of how to manipulate them without using the cognitive
and conceptual resources of particular communities. In other words
(to put it crudely), in order to understand the world we must
simultaneously understand one another. In everyday life, in so far
as common sense is characteristically unexamined, we tend not to
notice this social aspect and imagine that we can know objects in an
unmediated fashion. In common sense, we think with our beliefs
and concepts but not about them.30
The other (interdependent) relationship in which the subject
stands—to the object—is also widely misunderstood in that it is
frequently conceived of as merely contemplative rather than
practical. It is therefore not a question of knowledge developing



autonomously first and then (perhaps) being applied in a practical
context later: knowledge and practice are tied from the start. (But
again, note how the common-sense set of dualisms makes it
difficult to see this.) Even ‘pure’ science is also a set of practices.
The importance and interdependence of these two dimensions of
knowledge can be readily appreciated by recalling experiences of
learning a new skill or science. For instance, in mineralogy, it can
take weeks to begin to understand the concepts and to learn how to
look at the images under the microscope so that we see particular
minerals rather than pretty kaleidoscope patterns. And we achieve
this not just by looking but by doing things with the minerals and
microscope. For a while we may feel lost because the two
dimensions do not ‘connect up’; in using the instruments and
materials we seem only to be ‘going through the motions’ without
knowing why, while using the concepts feels like merely ‘mouthing’
or ‘parroting’ without understanding them. Later, connecting up
the two dimensions becomes ‘second nature’ and we are then
tempted to forget the dual relationship in which we stand as
subjects so that we may imagine that we have acquired a ‘stock of
knowledge’ without either material work or communicative
interaction.


If we broaden the meaning of ‘practice’ to include both these
dimensions, it can be seen that the nature of the practice both
determines and is determined by the kind of subject and object
which it links. For example, a cook and a nutritionist, or an
accountant and an economist have certain interests in common, yet
they are different kinds of ‘subject’ with differently defined objects,
the differences being determined by their practices, in terms of the
types of conceptual tools they use and material actions and social
relations in which they engage. Yet it is still common to compare
knowledge in different communities and at different points in
history in abstraction from these practical contexts as if they were
merely different modes of contemplating the world.
Although these two aspects of practice are interdependent, they
are, as noted above, qualitatively different. In Figure 2, the crucial
aspect of the social relations between subjects is the sharing of
meaning. In the case of knowledge of non-social objects the
relationship between 3 and O is not itself social. Even though it
requires the application of concepts and a language which can only
be gained in a social context, the object itself does not include
concepts or meanings.31 Non-social phenomena are impervious to
the meanings we attach to them. Although one could say that such
objects are ‘socially-defined’, they are not socially-produced.
Definition and production are utterly different, though some of the
literature which has stressed the idea of ‘the social construction of
reality’ tends to forget this, as if when we abandoned the flat earth
theory for a spherical earth theory, the earth itself changed shape!32
‘Subjects’, however, interact on the basis of shared understandings
which can be changed. Nature can be altered but through work and
not merely by changing systems of meaning: non-social objects such
as atoms do not act on the basis of shared understandings and so
are not susceptible to change in them. This may seem all very
obvious, but it is surprising how often change on the left side of the
diagram (conceptual change) is confused with change on the right.
On the other hand, given that it is only via the left side that we can
make sense of the right, perhaps it isn’t so surprising!
What does the relationship look like where the object is society?
(Note, once again, that I do not at this stage want to restrict the
discussion to ‘scientific study’.) In so far as this object includes
other subjects and their interaction, then the relationship should
have some features in common with that between the subjects on
the left side, so that the diagram becomes symmetrical (Figure 3).


For expositional clarity, the diagram shows two separate
language communities, which might represent situations such as
those found in history or the study of other cultures. It is, of course,
more common for 3 and O to be in the same language community
or society. Given that even anthropological or historical
investigation requires the establishment of conceptual connections
between the two communities, the separation in the diagram should
perhaps be regarded as an analytical device rather than a widely
applicable substantive description. In practice, there is usually a
partial identity of subject and object,33 so that we are often already
familiar with the meaning of the social phenomena in our ‘object’.
Nevertheless, even where the identity is full rather than partial, it is
possible for the subject 3 to characterize Os’ knowledge as wrong
or incomplete, and vice versa. Given the equivalence of the
horizontal subject-object relationship in Figure 3 to those within
language communities, social knowledge, including social science,
is sometimes said to stand in a ‘dialogic’ relationship with its object,
or in a subject-subject relation rather than a subject-object relation.
As we shall see, this relationship is widely misunderstood and needs
careful analysis, but before embarking on this, there still remain
some further modifications to be made to the diagram.



Understanding social phenomena is by no means just a question
of understanding concepts in society and the meanings of
practices.34 In the study of the British economy, for example, we
need to know not just what, say, ‘monetarism’ or ‘inflation
accounting’ mean to those who have claimed to put them into
practice; we also need to know under what conditions, to what
extent and with what effects they have been used. Social
phenomena have a crucial material dimension and are closely
associated everywhere to relationships with nature, both in its
virgin and its artificially transformed states. Knowledge of society,
whether scientific or lay, should therefore always include reference
to this material side, although it tends to be overlooked in some
‘interpretive’ approaches to sociology and anthropology (Figure 4).
It will be noted that the lines relating the communities to nature
correspond to the horizontal subject-object relations in Figure 2. As
such these involve a material, practical relationship. However, the
situation in social science is more complex for two reasons: 1 the
unavailability of experiments makes it more difficult to use such



material interventions for scientific purposes;35 2 social phenomena
can be changed intrinsically by learning and adjusting to the
subject’s understanding. It is not just that social experiments may
be deemed undesirable, it is also that social phenomena are likely to
be irreversibly changed by them in a way which does not happen
with non-social phenomena, which learn nothing from being
manipulated. In the desire to know society as it is, rather than as it
might be when modified by responding to our investigations under
uncontrolled conditions, it has widely been assumed that social
science should try to neutralize such interactive effects. As we shall
see, this position is being increasingly challenged—with important
implications for the role of social science in society. But for now, it
can at least be noted that characteristic 1 does not automatically
reduce social science’s relationship with its object to a purely
contemplative one, precisely because of 2.

Some implications of subject-object relations

In some ways the above account may seem too obvious to warrant
such laborious treatment. Yet the implications, particularly of
Figures 3 and 4, are profoundly at odds with the dominant
conceptual framework of oppositions of ‘subjective and objective’,
‘thought and action’, etc., in which we are accustomed to think (see
above p. 25). Failure to grasp these implications underlies some of
the most common misunderstandings of social science, but
unfortunately the failure is as common in social science itself as it
is in natural science and everyday knowledge. Given their extent, it
is necessary to proceed rather slowly and carefully in examining
what is implied by these last two diagrams.
The first point concerns the ‘intrinsically-meaningful’ or
‘concept-dependent’ nature of social phenomena.36 What does this
mean? It obviously denies the (tempting) assumption that meanings
are merely descriptions which are only externally applied to social
phenomena, as they are to non-social objects. The correct point
that ideas and meanings are not the same as material objects lends
some support to the ‘mental-material’ and ‘subjective-objective’
dualisms. Yet this type of thinking also makes it difficult to see how
the material structure of society—its institutions, social relations
and artefacts—are dependent on social meanings in various ways.
The most obvious candidates for intrinsically meaningful social


phenomena are the ideas, beliefs, concepts and knowledge held by
people in society. As part of the object—as well as the subject—of
knowledge, their meaning must be understood. There is no
equivalent of this where non-social phenomena are concerned. As
will be shown, this distinction (embodied in the contrast between
Figures 2 and 3) constitutes an absolutely fundamental difference
between social science, the humanities and everyday social
knowledge on the one hand and informal and scientific knowledge
on the other. In studying a fascist society we must interpret what
fascism means in it, for its members. The same goes for social
‘objects’ such as status, politics, nationality and gender, to name but
a few: but it does not apply to objects such as atoms, cells, black
holes or rock formations.
As we have seen, the point that these ideas and meanings are
not only in society but about society tempts us back into the
common-sense framework—back into the separation of
knowledge, language and meaning from the world of objects.
Against this, the crucial point to remember is that social
phenomena are concept-dependent. Unlike natural (i.e. non-social
objects) they are not impervious to the meanings ascribed to them.
What the practices, institutions, rules, roles or relationships are
depends on what they mean in society to its members. In one of
the most influential discussions of the constitutive role of meaning
in society, the philosopher Peter Winch has argued that the
essential feature of social institutions is that individuals have a
practical knowledge of more or less tacit constitutive rules
concerning not only what can and cannot be done but how things
should be done.37 Nevertheless, the influence of the common-
sense oppositions or dualisms mentioned above is such that this
argument tends to produce bafflement or resistance, so I will
illustrate it with several examples.
Money, and the institutions and practices associated with it, are
extremely important in our society (‘money makes the world go
round!’). A necessary condition of the use of money is that users
should have some understanding of what the act of exchanging
little metal discs and specially printed pieces of paper for
commodities means or ‘stands for’. The users must have some
concept of money and also of related phenomena such as rights of
ownership, exchange, etc. Hence these social phenomena are
‘concept-dependent’.
Likewise, for conversations, interviews, seminars or debates to


take place, the participants must have a practical knowledge of the
rules concerning what is supposed to happen in such situations.
A third and rather well-worn example of concept-dependent
practices is that of voting and holding elections. A necessary
condition for the holding of elections is that people must have some
understanding of what elections, voting, ballot papers, candidates,
democracy and so on mean. If we forced uncomprehending
individuals to mark crosses beside names on ballot papers, it would
not count as a proper election. Finally, given the symmetry of
Figure 3 we can treat social science itself as an example of an
intrinsically meaningful practice.
In all these cases and a host of others we can distinguish between
the physical ‘behaviour’ and the meaning of the ‘actions’ involved in
the practices. In the case of using money, we could observe the
physical behaviour of handing over the little metal discs until the
cows came home and we could use every statistical technique in the
book to process our observational data, yet if we didn’t know the
meanings on which the use of money is dependent in the society
under study, we would still not have any idea of what was actually
happening, or what kind of’action’ it was. Accordingly, Winch and
others have argued that this kind of understanding requires not the
amassing of empirical data but a conceptual or philosophical analysis
of the action and the rules implicit in it.38 ‘Mere’ physical behaviour
such as blinking, walking, sleeping or swallowing has no intrinsic
meaning, although in exceptional circumstances some of these can
acquire a certain social significance—for example, the disapproving
cough. Many actions are conventionally associated with physical
behaviour, but some are not; examples of the latter case are
remaining silent under interrogation or deciding not to vote.
Sometimes the same behaviour can, in different contexts,
constitute different meaningful actions. The physical behaviour of
different political groups in demonstrations may be very similar, yet
the meaning of their actions could be utterly different. I may raise
my hand in a meeting, but whether this constitutes voting, asking to
speak or bidding in an auction depends on the context and what the
other ‘social actors’ take it to mean.
Note that by ‘constitutive meanings’ or ‘concepts in society’ I
most emphatically do not mean simply the subjective beliefs,
opinions or attitudes of individuals. This conflation follows readily
from the conceptual framework of dualisms discussed earlier.
Those trapped within it tend to react to the above arguments by


assuming that constitutive meanings in society are nothing more
than the subjective beliefs of individuals which can be ascertained
through questionnaires or interviews and then treated as
untroublesome objective facts about those individuals. Meaning, on
this common-sense account, is reduced to either ‘private’, subjective
‘feelings’ or opinions—expressions of Inner states’—or references
to things. What is missing in this conceptual framework is any
recognition of the properties of language mentioned earlier. Nor
has it any concept of meaning as being for a subject, for a person,
or of utterances and actions meaning something to someone.39
Moreover, and related to this, there is a lack of recognition of the
intersubjective context of language: to speak or write is to enter
into a social relationship.40 As was explained in our earlier remarks
about the contexts of knowledge, even our most personal feelings
or opinions can only be constructed and communicated (and hence
have any chance of becoming constitutive or having any impression
or influence on others) within intersubjectively-understood (though
often non-verbal) terms. Although they do not realize it, those who
would reduce the interpretation of meaning to an opinion (or
belief) data-gathering exercise can only make sense of their data by
already presupposing knowledge of the meanings of the vocabulary
in which they are constructed. It is not merely that beliefs are
shaped by others, but that they are constructed in terms of
intersubjectively-available meanings.
Likewise social practice does not consist in the collisions of
individuals acting out their private beliefs, using language only as a
set of labels for their feelings (expressive function) or for the states
of the outside world (prepositional function). As has been argued,
language has a social function through which actions are co-
ordinated (or opposed) and people communicate with one another.
Beliefs and opinions are not the only phenomena which are
borne by individuals and yet are socially constituted. Roles and
personal identities also generally cannot be determined
unilaterally by individuals (or even by groups sometimes). You
cannot simply become an employed person by believing and
declaring yourself to be one. Whether you can become one
depends on (among other things) what other people are prepared
to take you as and on what they themselves have become (e.g.
whether they control access to the means of production).
Intersubjectivity is therefore an essential category for
understanding not only how scientists and others gain knowledge


of the social world (the epistemological relation) but also how
societies themselves cohere and function.
Material arrangements are also important in the determination
and confirmation of the meaning of practices within societies.
Consider the example of the concepts ‘public’ and ’private‘.
Although their meanings have certainly not been static, they have
informed actions in our society for centuries and have in turn been
objectified in its material organization, most obviously and simply
in the enclosed and locked spaces which are interpreted as
confirming the conceptual distinctions on which the actions
producing the material arrangements depend.
Sometimes material objects which do not depend at all for their
existence upon our conception of them may nevertheless be
ascribed a concept-dependent (symbolic) function in society.
Obvious examples are gold and diamonds. Manufactured objects
such as gold coins or fast cars are constructed out of intrinsically
meaningless objects, but signify certain concepts in their design, use
and function. The fast car not only objectifies technical knowledge
but also acts as a bearer of macho social imagery. Male owners of
such objects assume that others will respond in ways which confirm
their self-image, though, of course, they may inadvertently prompt
a debunking. The point to be made here is that although, in one
sense, material objects are intrinsically meaningless, their use and
functioning in society is concept-dependent. Conversely, although
systems of meanings and beliefs are not themselves material, they
usually require some material mode of objectification if they are to
communicate and function socially in a stable manner. In other
words, practices, material constructions and systems of meanings
are reciprocally confirming.41
Given this ‘reciprocal confirmation’, we usually find that
changes in meanings and practices go hand in hand. The struggle of
feminists and anti-racists to erase the negative meanings associated
with women and blacks cannot be effective purely at the level of
semantic battles. It must also involve the dislocation of those
material arrangements which objectively restrict them (e.g. access
to paid work) and those which as a matter of convention are
interpreted by sexists and racists as reciprocally confirming these
negative meanings. Understanding concepts in society and how
they change therefore requires an understanding of the material
practices associated with them and the way in which they are
contested. As Bourdieu puts it, unquestioning use of everyday


categories for things such as occupations or ethnic groups amounts
to ‘settling on paper issues that are not settled in reality, where they
are the stake of ongoing struggle’.42
A common reaction to these claims is to concede them but then
assume that they are only relevant for understanding small-scale
features of the social world, e.g. the way in which interpersonal
relations are reproduced. While it is true that most social scientists
who have made this process of reciprocal confirmation of meaning
and practice their specialism have concentrated on micro-
phenomena, large-scale phenomena such as the reproduction of
status systems, forms of political organization, nationalism and
religious systems are no less concept-dependent. 43 Raymond
Williams’s studies of shifts in social concepts and practices such as
‘democracy’, ‘Individualism’, ‘art’, ‘culture’ and ‘Industry’, in
Culture and Society illustrate this point.44 (The fact that many
social scientists don’t consider this as social science is indicative of
the ‘scientism’ and widespread ignorance of the significance of
constitutive meanings.)
There is, of course, another kind of dependence between the
realms of ideas and matter, which derives from the fact that people
are themselves material, animal and part of nature such that they
are subject to certain of its causal laws and conditions. Whichever
system of meanings societies adopt, they must satisfy certain basic
material needs in order to survive. This might be called a materialist
principle but it is not the kind in which satisfaction of material
needs must chronologically precede communication, culture, etc.,
for even the most basic and desperately needed material
requirements are simultaneously interpreted in terms of some kind
of system of meanings.45
So nothing I have said about the reciprocal relationship between
the construction of meaning and constructions and use of material
environments is incompatible with the ‘materialist principle’ thus
qualified. Unfortunately, ‘vulgar materialists’ often forget the
former relationship while students of the construction of meaning
(‘vulgar symbolic interactionists’?) often forget the latter. Social
beings live neither on bread alone nor on ideas and symbols alone.
Systems of domination invariably exploit both types of
dependence. They are maintained not only through the
appropriation, control and allocation of essential material
requirements by the dominant class, race or gender, but also
through the reproduction of particular systems of meanings which


support them.46 The relevant constitutive meanings (e.g. concerning
what it is to be a boss, master-race, untouchable, husband or wife)
are certainly not neutral or indifferent to their associated practices
and different groups have very different or even contradictory
material stakes in their reproduction or transformation.
I hope that the arguments and examples of the last few pages
have demonstrated that the initially apparently obvious claims
about subject-object relations and the context of knowledge have
implications which go beyond the conduct of social science to social
practice in general.

Verstehen

Having discussed what the ‘concept-dependence of social
phenomena’ means, I will now look more closely at the kind of
understanding involved. It is emphasized that the understanding
referred to here is common to all the relationships shown in Figure
3: it is not unique to social science, and the relationship between S
and Os (subject and social object). Any member of a society
achieves this understanding in everyday life; indeed it is precisely
because it is universal that it is often not noticed.
The discipline or science concerned with the interpretation of
meaning is called ‘hermeneutics’. Using this term we can say that
the study of natural objects (Figure 2) only involves a ‘single her-
meneutic’ (S1, S2 . . . ,Sn) while the study of ideas and concept-
dependent social phenomena involves a ‘double hermeneutic’.47
It is sometimes said of someone that they ‘read’ a social situation
well or badly. This is a revealing description, for the understanding
to which we refer, sometimes termed ‘verstehen’, is rather like that
used in and obtained from reading a book.48 We do not understand
a book (any more than we come to understand a foreign language)
by observing and analysing the shape of words or their frequency of
occurrence, but by interpreting their meaning. To this reading, we
always bring interpretive skills and some kind of pre-understanding
(though not necessarily a correct one) of what the text might be
about. In other words there is an interpenetration and engagement
of the ‘frames of meaning’ of the reader and the text. We cannot
approach the text with an empty mind in the hope of understanding
it in an unmediated fashion, for our own frame of meaning is an
indispensable tool or resource for understanding.49


However, the role of meaning in social interaction in everyday
life is usually different from that in a discourse, such as a text or an
argument, in that many of the successive elements of the
interactions in the former do not relate to one another in a logical
and conceptually consistent way. For example, in a confrontation
between two nations, although conflict requires communicative
interaction, responses are unlikely to succeed one another logically,
as if they were governed merely by the force of the better argument;
they are more likely to be determined by relative economic
strength, membership of power blocs, or contingencies such as
unanticipated consequences of political changes within each
country. Particularly where actors state their intentions fairly
formally, we should be wary of assuming that what appears to be
coherent on paper will be possible in practice; political manifestos
provide a good illustration of the danger! The analogy with reading
a text is useful for distinguishing the situation from that of natural
science, but only up to a point. The ‘text’ of actual social processes
is usually highly disjointed and often contradictory, and whereas it
is not generally necessary to know how a book was produced in
order to understand it, little sense can be made of social
interactions like international conflicts without exploring the
production of particular actions.50
As Figure 4 showed, hermeneutics is not the only kind of
understanding used in social science or everyday social practice. Yet
it is certainly the most widely misunderstood. I shall therefore
attempt to counter some of the misconceptions and objections.51
Perhaps the most common misunderstanding runs like this:
‘Social science has to concern itself with the subjective as well as the
objective, with people’s opinions and feelings as well as their
material states and circumstances. Understanding why people act
as they do requires that we examine this subjective side and for this
we need to “empathize” with them, by asking ourselves what we
would have done in their circumstances.’ Note again how the
subjective-objective dualism is asserted and intersubjective
meanings are collapsed back into subjective, essentially private,
opinions and feelings. Once this adulterated account of the
hermeneutic element of social knowledge has been taken as
authoritative, it is open to certain typical objections. One is that
while empathy may be a useful source of hunches or hypotheses
about why actions occur it is not a privileged source and what
matters is not where such explanatory hypotheses come from but


how they stand up to test. As one critic put it: ‘Empathy,
understanding and the like may help the researcher, but it enters
into the system of statements as little as does a good cup of coffee
which helped the researcher do his work’.52 The absurdity of this
‘cup-of-coffee-theory-of-understanding’ is illustrated by one of the
most famous critics of verstehen, Abel, who gave as an example the
problem of explaining why the marriage rate changes from year to
year in a certain community.53 Verstehen is presented as the use of
empathy to understand the motives of actors and hence as a source
of hypotheses explaining their actions. Once it is reduced to this
role, verstehen can easily be relegated to a dispensable status. But
the absurdity derives from the fact that simply by already knowing
what marriage is—as an intrinsically meaningful social
phenomenon—Abel unwittingly presupposes verstehen, not as
empathy but as the understanding of constitutive meanings, just as
any person presupposes it in social action. Indeed, without
verstehen, Abel would not be a social actor.
Note also that this implies that verstehen is universal: it is not a
special technique or procedure but is common to all knowledge,
both of nature (where it is restricted to a single hermeneutic, as in
Figure 2) and of society (where it is situated in a double
hermeneutic, as in Figures 3 and 4). However, this is not to deny
that it is used differently according to context. The intellectual’s
interpretation of meaning is (or should be!) rigorous and self-
aware, thinking, as noted earlier, about beliefs and concepts as well
as with them. By contrast, a very much less examined kind of
interpretive understanding is used in everyday, practical contexts,
where people are rarely aware that their actions presuppose it. It is
exactly this unawareness which explains the above
misunderstanding of verstehen by unreflective social scientists. In
everyday practice, however, it must be admitted that too much self-
consciousness of the processes by which people achieve mutual
understanding can actually interfere with the successful execution
of the most mundane social acts, such as holding a conversation.
So, although verstehen is common to knowledge in any context, it
does not take the same form in each.
Another common misconception about verstehen is the
assumption that understanding implies agreement.54 Once this is
accepted, it is, of course, difficult to make sense of conflict and
disagreement in society. However, to say that social actions and
communication take place on the basis of common understandings

andrew.sayer(1992)method.in.social.science a.realist.approach.2ed.ed

andrew.sayer(1992)method.in.social.science a.realist.approach.2ed.ed

Recomendados

Recomendados

Mais conteúdo relacionado

Mais procurados

Mais procurados (20)

Destaque

Destaque (6)

Semelhante a andrew.sayer(1992)method.in.social.science a.realist.approach.2ed.ed

Semelhante a andrew.sayer(1992)method.in.social.science a.realist.approach.2ed.ed (20)

Mais de jane tsai

Mais de jane tsai (19)

Último

Último (20)

andrew.sayer(1992)method.in.social.science a.realist.approach.2ed.ed